ABSTRACT Melanoma of the eye's uveal tract is fatal in approximately half of patients because of the development of metastatic disease that is invariably refractory to conventional cytotoxic treatments. Patients harbor subclinical micrometastases at presentation. The objective is to develop a treatment that does not produce cytotoxicity, but rather one that promotes terminal cellular differentiation of micrometastases. The central hypothesis is that aberrant epigenetic repression of differentiation genes by DNA methylation is a key process in uveal melanoma development. Decitabine reactivates genes suppressed by DNA methylation by depleting DNA methyl-transferase (DNMT). In preclinical models, frequent, intermittent, low-dose decitabine is highly effective in promoting differentiatio and terminating proliferation of uveal melanoma without producing toxicity. There are barriers to translating this observation to the clinic. The major barrier is pharmacologic. Cytidine deaminase (CDA) catalyzes rapid inactivation of decitabine. CDA also limits oral therapy, which produces the most favorable pharmacokinetics for depleting DNMT. Assessing melanoma differentiation clinically is difficult. Furthermore, drug effects are systemic, and host responses that can impact the anti-tumor effects can also result. A phase I clinical trial will be performed in patients with uveal melanoma of decitabine combined with tetrahydrouridine, a clinically applicable CDA inhibitor that can address pharmacologic resistance. Both drugs will be administered orally, not pulse-cycle but rather frequently and intermittently, and over a prolonged period. The aims are to establish a safe and biologically active dose and to evaluate mechanisms of regulation. Patients with high-risk tumor genotypes will be enrolled after primary therapy into three successive dosing cohorts consisting of six to ten patients. Safety will be established using standard clinical and laboratory toxicity assessments. The optimal biologic dose will be defined as the dose that results in a significant increase in blood differentiation biomarkers identified i preclinical studies. Regulation will be examined by quantifying DNMT levels and by measuring host factors, including immune mediators. The results of this clinical trial, which will be the firt to modify decitabine formulation, dose, and schedule for differentiation objectives, will improve the understanding of epigenetic regulation in uveal melanoma and should lead to more effective, less toxic therapy for uveal melanoma and potentially other refractory solid tumors. Novel blood biomarkers of melanoma differentiation and decitabine effects will also be developed.